DEVELOPING.md: add pprof instructions

Signed-off-by: Silvio Moioli <[email protected]>

DEVELOPING.md

@@ -256,6 +256,41 @@ That being said, downstream `fleet-agent` deployments can perform Kubernetes API
 While network traffic is major point of consideration, we also have to consider whether our performance issues are **compute-based**, **memory-based**, or **network-based**.
 For example: you may encounter a pod with high compute usage, but that could be caused by heightened network traffic received from the _truly_ malfunctioning pod.
 
+### Using pprof
+
+[http pprof](https://pkg.go.dev/net/http/pprof) handlers are enabled by default with all [default profiles](https://pkg.go.dev/runtime/pprof#Profile) under the `/debug/pprof` prefix.
+
+Additionally, it is possible to enable continuous CPU profiling for `fleetcontroller` to observe how CPU usage changes over time.
+
+Add the following extra Helm values:
+```yaml
+cpuPprof:
+  period: "60s"
+  volumeConfiguration:
+    hostPath:
+      path: /tmp/pprof
+      type: DirectoryOrCreate
+```
+
+Notes:
+ - `period` is the pprof CPU period and can be changed with any other value
+ - `volumeConfiguration` can be any valid volume configuration (not necessarily `hostPath`)
+
+Alternatively, use the following Helm commandline arguments:
+```
+--set cpuPprof.period=60s \
+--set cpuPprof.volumeConfiguration.hostPath.path=/tmp/pprof \
+--set cpuPprof.volumeConfiguration.hostPath.type=DirectoryOrCreate \
+```
+
+If using `hostPath`, make sure that the target directory (`/tmp/pprof` in above examples) is writable.
+
+Profiles can be inspected with the [pprof tool](https://github.com/google/pprof), eg.:
+
+```sh
+pprof -http=localhost:5000 ./2022-11-04_19_47_18.pprof.fleetcontroller.samples.cpu.pb.gz
+```
+
 ### Tips and Tricks
 
 Since the Fleet components' controller framework of choice is [Wrangler](https://github.com/rancher/wrangler), we can share caches and avoid unnecessary API requests.